I have enough evidence for a dense, publication-ready investigation. Here is the report.

The Volunteers Who Decide a Child's Fate: How CASA's Unregulated Advocates Move Custody Cases With 30 Hours of Training and No Accountability

In dependency courtrooms across all 50 states, some of the most consequential recommendations a judge will hear — whether a child goes home, stays in foster care, is placed with relatives, or is freed for adoption — come not from a licensed social worker, a psychologist, or an attorney, but from a volunteer with roughly 30 hours of training and no professional credential of any kind. These are Court Appointed Special Advocates (CASAs) and volunteer guardians ad litem, and judges follow their recommendations in the overwhelming majority of cases. Yet the best national evaluation ever commissioned by CASA's own leadership found that children assigned a CASA were more likely to be removed from their parents and less likely to be reunified or placed with kin, that volunteers spent strikingly little time on the cases of Black children, and that the volunteer corps is roughly 80–90 percent white while serving families that are overwhelmingly poor and disproportionately non-white. A generation of researchers has concluded there is not enough rigorous evidence to call CASA an evidence-based practice at all — even as the program has drawn tens of millions of federal dollars a year through the Victims of Crime Act and related child-abuse statutes, and even as the advocates themselves are shielded from lawsuits by broad judicial immunity. This is an investigation into an institution that exercises quasi-governmental power over families with almost none of the checks that normally accompany it.

How the Machine Actually Works

The CASA model is deceptively simple, and that simplicity is the problem. A juvenile or family court judge appoints a volunteer to a child who is the subject of an abuse or neglect (dependency) proceeding. The volunteer is expected to interview the child, the parents, foster parents, teachers, and caseworkers; review records; and then submit a written report to the court with a recommendation about where the child should live and what should happen to the family. Under the widely adopted "best interests of the child" standard, that recommendation is filtered through the personal judgment of the volunteer — a standard that legal scholars have long criticized as so elastic it functions as a mirror for the decision-maker's own values and biases.

The training threshold is minimal. California's Rule of Court 5.655, typical of the national standard set by the National CASA/GAL Association, requires only a minimum of 30 hours of pre-service instruction before a volunteer is sworn in and assigned to a real family in crisis (Judicial Branch of California). By comparison, a licensed cosmetologist in most states needs well over 1,000 hours. A CASA needs no degree in social work, no background in child development, no legal education, and no clinical supervision by a licensed professional. There is no national competency examination, no licensure, and no mandatory continuing-education body that can revoke a credential — because there is no credential to revoke.

The advocate's influence, however, is anything but minimal. National CASA's own promotional materials and independent reporting agree that judges "often do exactly what the advocates recommend" (Youth Today). Some jurisdictions blur the line between advocate and legal party entirely: in many states the volunteer serves as, or works alongside, the guardian ad litem — a role that can carry the weight of the child's legal representative even though the person filling it may have no law degree. The result is a structure in which an untrained, unpaid, unregulated volunteer can shape a binding custody outcome, and a judge with a crushing docket has every incentive to defer to the one person in the room who claims to have spent the most time with the child.

The Data Its Own Backers Buried

The most damning evidence against CASA did not come from critics. It came from Caliber Associates, the research firm National CASA itself commissioned to conduct a national evaluation, released in 2004. The findings were so unfavorable that they became a lasting embarrassment for the movement. According to the evaluation and subsequent reporting, children assigned a CASA were removed from their parents at dramatically higher rates than children without one — with reporting citing removal in roughly 89 percent of CASA cases versus about 18 percent of non-CASA cases — and CASA children were less likely to be reunified with their parents and less likely to be placed with kin (Youth Today).

Just as alarming was what the study revealed about race. The evaluation found that volunteers — an overwhelmingly white and female group — spent little time on cases overall and even less time on the cases of Black children (Youth Today). In a system where a recommendation is only as good as the investigation behind it, spending less time on a Black child's case while still holding the pen on that child's future is not a neutral act.

The pattern has held up in more recent, peer-reviewed work. A 2020 study published in Child Maltreatment by Cynthia Osborne and colleagues, "The Effect of CASA on Child Welfare Permanency Outcomes," found that children with a CASA had lower odds of being reunified with their parents, greater odds of being adopted if not reunified, and lower odds of being placed in permanent kin guardianship than children who were never assigned a CASA (SAGE / Child Maltreatment; PubMed). In plain terms: two decades after Caliber, the association between having a CASA and being permanently separated from one's family — rather than returned to it — persisted.

Meanwhile, the broader scientific verdict is that CASA simply has not been proven to work. A Campbell Collaboration systematic review protocol and subsequent scholarship concluded that, because of equivocal findings and pervasive methodological weaknesses — most notably selection bias — there is not enough evidence to establish CASA as an evidence-based practice (Campbell Systematic Reviews; eScholarship). No randomized controlled trial has ever demonstrated that CASA involvement improves child safety. The favorable-sounding claims the program markets — that CASA kids spend less time in foster care, or that courts get "higher quality information" — come from studies that cannot rule out the possibility that judges simply assign CASAs to different kinds of cases in the first place. After more than four decades and hundreds of millions in public funding, the core question — does this make children safer? — remains unanswered by the standard we apply to a new medication or a new curriculum.

The Money

CASA is not a grassroots charity operating on bake sales. It is a federally subsidized network. Under the Victims of Crime Act (VOCA), state and local CASA/GAL programs are eligible for VOCA victim-assistance grants, and in 2020 more than 500 state and local CASA/GAL programs across 40 states received a total of over $83 million in VOCA funding, with more than ten states channeling over $1 million each to CASA organizations (National CASA/GAL Association). VOCA dollars are financed not by taxpayers directly but by federal criminal fines and penalties deposited in the Crime Victims Fund — money whose disbursement carries its own administrative provisions under Chapter 201 of Title 34 (including 34 U.S.C. § 20110).

At the top of the pyramid sits the National CASA/GAL Association for Children, which administers a national sub-grant pipeline under a cooperative agreement with the Office of Juvenile Justice and Delinquency Prevention (OJJDP), authorized by the Victims of Child Abuse Act, 34 U.S.C. § 20323 (the OJJDP FY2021 VOCA Children's Advocacy National Subgrant Program, CFDA #16.756) (National CASA/GAL Association). That federal relationship, essentially uninterrupted since 1994, is the financial spine of the movement: National CASA takes federal money and re-grants it to state and local chapters, giving it enormous leverage over standards, training, and messaging across the entire network.

The scale of that dependence became vividly clear in the spring of 2025. On April 22, 2025, the U.S. Department of Justice issued more than 360 grant-termination notices, and National CASA/GAL was among the recipients; the government's stated rationale was that the awards "no longer effectuate the program goals or agency priorities." The termination cut off roughly $25 million in ongoing federal support (with some accounts of the total exposure ranging higher), forcing National CASA to suspend all OJJDP-funded services on April 24, 2025, terminate current grants to 101 state and local CASA/GAL organizations, and cancel plans to award 275 additional grants (The Imprint; CASA MN). That a single agency's decision could instantly freeze a nationwide "independent" child-advocacy network is itself a measure of how thoroughly the model runs on federal money.

The reinstatement is where the story turns revealing. National CASA appealed and won — but the restored grants came with strings that exposed how the organization had been framing race and bias in its own training. To comply, CASA leadership instructed local programs beginning May 15, 2025 to "Skip or Adapt as Directed" dozens of sentences from slide decks, volunteer manuals, and facilitator guides. Among the items volunteers were told to skip: a self-evaluation checklist asking whether each volunteer "understands and demonstrates self-awareness to eliminate the influence of personal biases and values when working with diverse groups" (The Imprint). In other words, to keep the federal money flowing, the nation's premier volunteer child-advocacy organization agreed to stop explicitly training its overwhelmingly white volunteers to check their own biases before making life-altering recommendations about disproportionately non-white children.

The Named Players and the Incentive Structure

The intellectual case against CASA has crystallized around one document in particular. In 2016, Amy Mulzer, then a staff attorney and clinical instructor at Brooklyn Law School, and Tara Urs, an attorney with the King County (Washington) Department of Public Defense, published "However Kindly Intentioned: Structural Racism and Volunteer CASA Programs" in the CUNY Law Review (SSRN). Citing a 2014 national survey showing CASA volunteers to be roughly 81 percent white and 82 percent female, deployed into families that are overwhelmingly poor and disproportionately non-white, the authors wrote that "this exercise of white supremacy is out in the open, obvious, direct," and that allowing CASAs to speak for child-welfare-involved children "is to take the structural racism underlying the child welfare system and give it a seat at the table" (The Imprint; NCCPR Blog).

The demographic mismatch the article describes is not rhetorical. Independent analyses have documented volunteer corps that were roughly 90 percent white and 79 percent female serving a child population that was 48 percent white and 36 percent African-American or biracial, and at least one large urban family court was found to be appointing CASAs across racial and ethnic groups in a manner disproportionate to the actual racial distribution of its caseload — meaning even the threshold question of which children get this powerful advocate is skewed (Who gets a CASA? — PMC/NCBI).

The incentive structure compounds the demographic one. CASA volunteers, disproportionately affluent and white, frequently encounter poverty — an empty refrigerator, an unstable apartment, a missed appointment — and, lacking training to distinguish poverty from neglect, may read material deprivation as parental unfitness. Critics including the National Coalition for Child Protection Reform (NCCPR) argue that volunteers end up "policing and punishing families for being poor and Black or Native American," recommending removal and foster care where a professional might have recommended the concrete services that would let a family stay together (UCLA Blueprint). The volunteer bears no cost for a wrong recommendation; the family bears all of it.

The Accountability Gap

Here is the structural core of the problem: the people wielding this power are, in most jurisdictions, legally untouchable. Courts have repeatedly held that guardians ad litem — the role many CASAs occupy — enjoy absolute or quasi-judicial immunity for acts within their court-appointed duties, because they are treated as an "arm of the court" performing judicial functions (ABA Journal). One appellate court acknowledged that a guardian ad litem's conduct — failing to investigate abuse allegations, declining to speak to a child's therapist, making a recommendation not in the child's best interest — "could be characterized as negligent and even reckless," and then held the guardian immune from suit anyway (The Daily Record). A handful of states have carved out exceptions — the Pennsylvania Supreme Court, for instance, held that a guardian ad litem is not immune from a legal-malpractice claim (Marshall Dennehey) — but the dominant rule leaves families with no meaningful recourse when a volunteer's recommendation is negligent, biased, or simply wrong.

Now stack the immunity on top of everything else that is missing. There is no national competency standard. There is no uniform conflict-of-interest regime with teeth beyond each local program's internal policies. There is no licensing board to discipline a volunteer who overreaches. There is no outcome data demonstrating the program works, because no randomized trial has ever been run. And the one entity positioned to demand rigor — the federal government funding the enterprise through OJJDP and VOCA — has historically underwritten the network without conditioning the money on proof of effectiveness or on measurable safety outcomes. The people who are supposed to be watching are the same people writing the checks, and until 2025 the checks flowed with remarkably few strings attached to results.

Local programs will correctly note that they have written ethics and confidentiality policies, that volunteers are prohibited from serving on cases involving relatives, and that ex parte communication with the judge is a serious policy violation (Volunteer Policy Manual, NY Courts). But these are internal rules enforced by the very organizations whose funding depends on placing volunteers — not external, enforceable standards backed by a licensing authority or a right of action for a wronged family. Self-regulation by a federally funded advocacy network is not the same thing as accountability.

Why It Matters — and What Would Fix It

The stakes are not abstract. Every recommendation that pushes a child toward removal and away from reunification or kinship placement carries lifelong consequences: children who "age out" of foster care face elevated risks of homelessness, incarceration, unemployment, early pregnancy, and trafficking. When the research repeatedly associates CASA involvement with less reunification, less kinship care, and more time in the foster system — the very outcomes child-welfare policy is supposed to minimize — the possibility that a well-meaning volunteer program is nudging children away from their families, disproportionately families of color, demands to be taken seriously rather than waved away with testimonials.

None of this requires abolishing the impulse behind CASA. Volunteers who care about children are a genuine asset. The problem is a governance vacuum around a function that is quasi-governmental in effect. A serious reform agenda would include: (1) a mandatory national competency and training standard far exceeding 30 hours, developed with input from child-development and social-work professionals, with a real credentialing body; (2) required racial-bias and poverty-versus-neglect training — the exact content the 2025 federal DEI purge stripped out — restored and made non-negotiable, because a corps that is 80-plus percent white deciding the fate of a caseload that is disproportionately Black and Native cannot responsibly skip it; (3) conditioning federal VOCA and OJJDP dollars on the collection and public reporting of outcome data, so that after 40 years the field can finally answer whether CASA improves child safety; (4) narrowing the immunity doctrine so that families harmed by a negligent or biased recommendation have some avenue of redress; and (5) enforceable, externally audited conflict-of-interest and equity-of-appointment standards, so that which children receive an advocate is not itself a function of a volunteer pool's comfort level.

Until those guardrails exist, the reality remains this: in thousands of courtrooms, the single most influential voice on whether a child goes home belongs to a person the system trained for a long weekend, cannot hold liable, never tested for competency, and has never proven makes children any safer. The families on the receiving end — poorer, browner, and more powerless than the volunteers deciding their futures — are the ones absorbing the cost of that experiment. A child-welfare system that claims to act in the "best interests of the child" owes those children more than good intentions and immunity from the consequences.

Sources: Youth Today — evaluation controversy · Osborne et al., Child Maltreatment (2020) · Campbell Systematic Review protocol · Establishing CASA as Evidence-Based (eScholarship) · National CASA — VOCA funding · The Imprint — funding terminated · The Imprint — grants restored after DEI removal · Mulzer & Urs, "However Kindly Intentioned" (SSRN) · The Imprint — "exercise of white supremacy" · Who Gets a CASA? (NCBI/PMC) · UCLA Blueprint — The Case of CASA · ABA Journal — GAL absolute immunity · The Daily Record — GAL qualified immunity · Judicial Branch of California — Rule 5.655 · NCCPR — The Case Against CASA